More generally, if one wants to learn about the distribution of opinions in a group X, one needs to make some kind
of a poll,
Unfortunately, it’s even worse than that, because the same issue (selection bias) arises in polls. In fact, a lot of missing data work that tries to deal with bias adjustment was done in the context of analysis of survey data, I think.
Unfortunately, it’s even worse than that, because the same issue (selection bias) arises in polls. In fact, a lot of missing data work that tries to deal with bias adjustment was done in the context of analysis of survey data, I think.